A clustering algorithm for mobile ad hoc networks based on spatial auto-correlation

Dynamicity of topology and lack of infrastructure in mobile ad hoc networks (MANETs) induce high control overhead for routing protocols and increase end-to-end delay as a result of complexity in mobility management. In this paper, the so-called WCA (Weighted Clustering Algorithm) is developed and a novel mobility measurement metric is introduced based on prediction of spatial auto-correlation between traversed paths of mobile users for clusterhead selection. Collaborative mobility behavior of nodes is represented by spatial auto-correlation index in order to find more stable clusters when the nodes are highly mobile and follow a dependent mobility model. We compare results of the proposed model with WCA and EWCA (Entropy-based WCA) in terms of reaffiliation frequency (rate of changes in cluster membership) to measure stability of clustered structure for different mobility scenarios. Simulation results show that our proposed algorithm improves performance of clustered structure for MANETs.